High-Frequency Interference Suppression. Interference is Identified and then Eliminated.
Abstract
This report investigates techniques for suppressing interferences that occur on a communication channel, specifically the high-frequency radio channel. The interferences are divided into 2 categories, those that occupy a narrow bandwidth with respect to the communicator's signal and those that occupy a bandwidth approximately equal to or greater than the communicator's bandwidth. The processing techniques which are examined are based upon using large time-bandwidth product (TW) signals, where TW is much larger than the symbol alphabet size, whereby large portions of the symbol TW space can be altered or totally eliminated while still maintaining quality communication due to the symbol coding redundancy. The suppression problem is divided into two areas; first, the identification of the interference and, second, its elimination. A simple threshold algorithm (called the quantile algorithm) is used for identification where a level is set based on one point of the exceedance function over the ensemble of the time-envelope samples (for broadband interference in the time domain) or the magnitude of frequency samples (for narrowband interference in the frequency domain). The suppression techniques investigated include excising (setting to zero) samples that exceed the threshold level, clipping at the threshold level, and a logarithmic weighting of samples. Simulations were used to evaluate various techniques and it was found that with only one type of interference present, the simple quantile algorithm and excision processing worked well for either narrowband or broadband interference.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 08, 1977
- Accession Number
- ADA049138
Entities
People
- Gary J. Brown